[
https://issues.apache.org/jira/browse/HDFS-7411?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14236549#comment-14236549
]
Ming Ma commented on HDFS-7411:
-------------------------------
Andrew, nice work. It appears I don't need to continue the work on
https://issues.apache.org/jira/browse/HDFS-7442. Some initial comments.
1. NN Memory impact on the additional decomNodeBlocks. It shouldn't be an issue
given admins won't decommission lots of nodes at the same time. But it might be
worth calling out some limit here. 100 nodes * 400k blocks per node * 8 bytes
per blockInfo reference, 320MB extra at the start of the decommission process?
2. It appears "dfs.namenode.decommission.blocks.per.node" description should
refer to "dfs.namenode.decommission.nodes.per.interval" instead.
3. It appears this patch also fixed
https://issues.apache.org/jira/browse/HDFS-5757 by calling decomNodeBlocks.put
during refreshNodes.
> Refactor and improve decommissioning logic into DecommissionManager
> -------------------------------------------------------------------
>
> Key: HDFS-7411
> URL: https://issues.apache.org/jira/browse/HDFS-7411
> Project: Hadoop HDFS
> Issue Type: Improvement
> Affects Versions: 2.5.1
> Reporter: Andrew Wang
> Assignee: Andrew Wang
> Attachments: hdfs-7411.001.patch, hdfs-7411.002.patch,
> hdfs-7411.003.patch
>
>
> Would be nice to split out decommission logic from DatanodeManager to
> DecommissionManager.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
